Face Recognition Enabled Door Access Control System

Abstract

Face recognition has found usage in many applications from the simplest systems to the more complex systems. The years of researches in this field has witness improvements in detection accuracy. However, few studies that incorporate face recognition for access control are suitable for restricted areas. This work aims at using face recognition for access control systems using Haar cascade feature descriptor and LBPH as classifier algorithm. The system was implemented on raspberry pi 3- model B board, which consists of a solenoid lock that is activated to grant access to the user when a matching face is detected by the classifier algorithm. In the event when there is no match between the detected face and the database facial images, a red light comes on and the doorbell rang. The accuracy, sensitivity, specificity, and precision of the LBPH algorithm are respectively 96%, 97.5%, 90.5%, and 97.5%. The delay in the processing of signals to activate the door lock and doorbell is not more than 4 seconds. Thus, the system is effective as a real-time access control mechanism. Keywords— Access control, Computer vision, Facial recognition, LBPH, OpenCV.